It's Breathtaking How Fast AI Is Screwing Up the Education System
gizmodo.com
gizmodo.com
Submitters: "Please submit the original source. If a post reports on something found on another site, submit the latter." - https://news.ycombinator.com/newsguidelines.html
I agree with all else but I get a different impression here.
The unethical behavior really ramped up not with participation trophies per se, but around the same time we started gamifying everything. People treated each other like NPCs in GTA to abuse for their own amusement or advancement. On the internet we stopped being people and became targets to destroy. Nothing appeals to psychopathic behavior like turning any environment into a feedback loop of reward-seeking.
"Gamifying" life should give anyone pause when you witness how people act within existing game systems. How much effort is spent policing fraud, abuse and antisocial behavior within zero-stakes environments of games themselves? Multi-player Minecraft is unplayable unless you band together into tribal cliques with private servers (de-facto ingroup "racism" with no inclusivity clause for trolls); otherwise random people will log in, destroy your world on purpose and leave. We see similar behavior in real life from terminally-online types committing arson and trying to bait law enforcement into killing people. It's not coincidental.
Gamifying anything drives a competitive environment in which people are compelled to win (dominate) by any means necessary. It's not ethical to force opponents into bankruptcy by ruthlessly exploiting them, but it's the literal point of Monopoly.
School itself is gamified through the reward of grades and privileges. The system you describe is the one we already have. People will always have incentive to cheat to get ahead, especially when competing for or trying to retain tenured positions.
Creativity is not a tap that you turn on and off. Sometimes it is hard to be creative, even if you enjoy it. Even if you are good at it.
When a student feels pressure to perform, and they have an LLM to fall back on, they can poison the well of their creativity. LLMs are that tap that they can always turn on.
Then soon enough they no longer want to put in the effort. Why, when an LLM can do it perfectly right now?
Assuming we get to the point where the output of these models is adequate for the task, I don't think that's a problem.
Sure, we'll end up with many fewer poets, playwrights, songwriters, photographers, and the like - but those we do have will be doing it because they're passionate about it. They'll be doing it in spite of the economics, not because of them.
I'm in my 40s, and very much a software engineer at this point. I don't want to do anything else professionally if I can help it. Yet, I have a Fujifilm X-Pro3 sitting on my desk right now. I've got about $5k in lenses in my bag, and still love photography. I've done it professionally. I know with certainty that I can sustain my family on it as a career, and I know how much work that is. I'm not interested.
Instead, I'm mostly the "official photographer" for my wife's side businesses. I take on jobs here and there for friends and acquaintances when they seem fun and interesting. I do just enough to keep my name out there in the community, so I can fall back to it if my "real job" goes away unexpectedly -- which is always a possibility working for startups!
Basically, I do think creative fields will shrink significantly. I agree that we'll see AI-generated art used more and more frequently. The quality will improve - though, honestly, there is already a market of almost unlimited size that wasn't being served by humans because the expected value of those works didn't justify their creation.
In fact, that's also a good point: a rise in AI-generated art usage does not necessarily mean a fall in human-created art. I think it will mean less human art and fewer human artists in time, but that may not be the case. It very well could be the case that most brands use AI, but those that want to set themselves apart as particularly high quality or luxury will lean more heavily on human artists than ever before.
The suckers who actually want to learn will get even more badly screwed in such a system unless the assessment is balanced to favour actual knowledge. And not only at final assessment but from start to end because if you get screwed in the short term because your low marks while you get to grips with the subject gets you penalised, it takes a lot of foresight and fortitude not to buckle and just spam for points. Doing it "right" requires more teacher engagement, more parental engagement and generally more autonomy, effort and money all round. And even then you have the ever present problem of which group do you spend each marginal unit if effort on: the stronger or weaker students?
That's not a misalignment. That's your "ambitious student" is being unwise and stupid. To this day, I have gaps in my knowledge and skills that I regret because I avoided things that I, in my childishness, did not consider "important" or "interesting." AI is going to make that far worse.
If AI makes their kids miscalculate and end up failing in a revamped AI-resistant educational system that actually requires learning to pass, I'll laugh.
Or, more likely, you're going to fail because you can't do the things you need to do to be successful in said careers. Or, at least, you're going to struggle because you're not very good at it.
I've met plenty of bad engineers, and let me tell you, their job seems a lot more stressful than mine. I'll pass.
I recall it being pretty obvious for a large number of years of my time in the education system which kids were cheating their way through, because they'd be able to parrot crap in writing or do perfectly on take-home things, but mysteriously be crap if you required them to reason about it in the moment.
I imagine we've all met people in our professional careers who could sell you a heater on the beach but can't reason their way out of a paper bag, as well - it's the same problem.
I think it very likely AI is the chisel which opened up one or many of the cracks but the cracks have existed for a very long time.
I think "what is the education system" is a great open-ended pub discussion topic.
The article makes a very good claim that for people with freedom to chose expensive ivy league education, actual knowledge acquisition is the last thing on many of their minds.
Brand new teachers design a lesson plan using two sentences and AI, students answer it using a screenshot and AI. I'm struggling to see who learned what there.
All we do is end up with teachers that don't know how to teach, and students that learn little.
Personally I feel there is a step change towards a poorer education, but others may not. I think it's worth debating this particular topic. Or tackle education as a whole...but that feels intractable.
We know that behavioral and cognitive patterns develop at a very young age, soon become ingrained, and usually affect the rest of our lives in one way or another. Kids who are never disciplined grow up to be entitled spoiled brats, so what do you think will happen to kids who habitually delegate any work that requires even modest cognitive effort to a black box?
I don't buy the argument that this is just another technical development that makes things easier and more efficient. The WWW made information easier to find than searching for books in a library and reading them, but fundamentally you still have to read the material, understand it, and apply it to your work. LLMs can just finish students' assignments for them and they don't even have to read it or spend spend any energy thinking about it.
for example: if you had repetitive busywork for homework assignments, it doesnt make sense to spend a lot of time on it, once you have proven to yourself that you know how to do it anyway.
But the second half of your post is kind of mean-spirited. "You and many others are not cut out for a broader education," is a rude, elitist way to call someone stupid. With few exceptions, I think practically anyone is "cut out" for higher education. What that education looks like is obviously different for each person, but that's as it should be, as the world is so vast that the human race can only somewhat learn a significant portion of it all if we all learn something a little (or a lot) different.
The university system is a special time and place that society has carved out where individuals can prioritize learning and experimenting above all other considerations, including (ideally) their job or paying bills. It's amazing that society has prioritized this kind of self-actualization, and "societal-actualization." It's maybe one of our greatest achievements! Anything that cheapens that experience is a shame, whether that is the belief that one's primary value is only equal to what they contribute to the economy, the outrageous costs of college tuition, or the allure to use ChatGPT and similar tools to net easy wins at the expense of building a wonderful mind.
It was not uncommon for a kindly, well-meaning Sister to slightly nudge a struggling student toward a correct answer, though it never happened to me personally. Yes that's right -- when you standardize testing, the teachers become the cheaters, because they have a vested interest in seeing students excel and driving up their institution's own rankings.
In fact my very recent student career was rather dismaying, because I encountered more than one teacher who was more than willing to cross a line and help a student into a better grade than what they could earn on their own. I didn't think it was fair, that I worked my ass off for grades and teachers would just basically join in the cheating. In fact, I tanked a grade in one class because I was incensed about the teacher's behavior. I blew up some important coursework and I directly, sternly told him not to fix up my grade. It was only a "B" but I was so pissed, at that point!
In the end, all our cards were sent off, literally to Iowa, I guess, and fed into giant A.I. machines and graded without human intervention. But our results were very human indeed. I have no reason to believe that our digital overlords failed to grade our S.A.T.s in the same manner, but the regime was slightly looser as we progressed through high school.
In a way, you can say that I lost my job to A.I. in 2024, because indeed I was tasked with grading student submissions, and by the time I was resigning about half of our workload was turned over to auto-testing multiple-choice quizzes and such that would be scored in the LMS itself.
While I personally rarely determined blatant A.I. usage to solve our homeworks, plagiarism was rampant and students were constantly copying from a very small pool of solved homeworks that were publicly posted. Some more savvy students paid a subscription for a cheating website that would provide even more content to them.
But indeed, if educators have relied so heavily on A.I. to produce and grade materials on our end, I suppose it was only a matter of time before turnabout is fair play.
I have major reservations about the evaluation of student essays solely based on an overgrown pile of matrices that has no comprehension of creativity nor literary technique.
A digital FIR filter is not intelligence, even though it can adapt to changing conditions.
My senior year of high school, I wanted to enroll at a community college as a concurrent student. My high school demanded that I stay until at least 1:45 every day; I wanted to take a class at 12:30, which would require me to leave high school before lunch at 11:45. They absolutely refused to budge on this.
I set up a meeting with the school principal and superintendent. I laid out my case for why I wanted to leave before lunch and tried to be persuasive. Once it was clear they weren't having it, I switched gears.
My senior class was 31 students. The mean scores on the state standardized test was used to determine funding status for the school and to identify schools to target for consolidation - and my school was on probation. If their scores fell more than a couple of points the state would revoke their accreditation and work toward merging them with a neighboring district.
On the other hand, I already had a college acceptance letter and a full scholarship. The test was completely irrelevant to me.
I looked my principal in the eye and told her that if she couldn't work with me, I understood, but "if you're not willing to work to help me meet my goals, why should I work to help you meet yours?" In short, I threatened to bomb the test if they couldn't find a way to let me leave campus for college classes at 11:45 every day.
To be clear, I wasn't saying I wouldn't try to do well; I was saying I was very confident that I could get a pure 0% on that test. It wouldn't impact my future in any way, but with 31 students in my class that 0 would mean a 1-2 percentile drop in the school's mean test scores for the year -- enough to go from "secure in funding for another year" to "50/50 chance the district would be dissolved".
I left every day that year at 11:45.
They didn't change the policy, but let me enroll in a math class that was far below my level. I got along well with the math teacher (who knew what was up) and she let me take the mid-term and final before class started so I never had to attend the class.
More importantly, it taught me to trust my kids, show them my perspective, and put my own interests on the line for them.
My dad was a teacher at that school; the people I threatened were his bosses. I talked to him before I did it - in fact, I had only made an offhand comment about being able to bomb that test in revenge. He's the one that pointed out that was leverage, and suggest that I consider explicitly using it.
I was terrified that he could have lost his job for that stunt. Much later, he told me that they never even brought it up to him. He retired from that school after outlasting both the principal and superintendent, so obviously it didn't hurt him that much.
As a result I make it a point to trust my kids when the stakes are bigger than they're comfortable with, and to highlight the potential effects when I think they may not be obvious to them.
It’s kinda hilarious that a lesser student would actually be less capable of such a threat as missing every question is slightly challenging (might pick the right answer by luck/error).
Parents evaluate the quality of private schools based on NAPLAN results.
It took me several seconds to recognize this acronym for a test I took decades ago but haven’t thought about since.
Wow, time to move them.
Lock them in a room for 4 hours to 6 hours with a list of carefully crafted questions to answer in writing (rather than ticking boxes) with no access to a computer.
See how AI will help them get out of that one.
And testing student progress on "home" work, give me a break. It has always been broken.
Way before AI existed, all you had to do is hire someone to do the work on your behalf, all the more easy if you're born in a family with money. Nothing new under the sun.
No. AI can be very helpful when trying to understand new topics. The author does not seem to understand that a technology is good or bad, nor in itself, but in how it is used.
The real question, which is totally unrelated to AI, is: "Why don't students care about their education?"
The point regarding technology is a given and already understood by any reasonable person. This is not our first encounter with disruptive "technology".
The germinal point is the confluence of behavioral patterns and influences that ultimately determine how technology is used.
OP: >Everybody who uses AI is going to get exponentially stupider, and the stupider they get, the more they’ll need to use AI to be able to do stuff that they were previously able to do with their minds.
This to me is a reasonably probable outcome. It is not the only possible outcome but our record with the internet (the 90s' disruptive technology with its own +/- possibilities) is the cautionary tale. Remember voices that very clearly predicted the surveillance society in the 90s?
[p.s. consider auto-correct for spelling. Would you agree with my impression that the generation that has grown up with this (hardly disruptive) tech no longer can spell words of their own language?]
So, going back to the agreed baseline of 'how we use tools' as the critical matter, it seems (based on historic facts) that we should take 'cassandras' seriously from the very onset of introducing disruptive technology at societal scale and take the necessary steps to insure that the incentive to use a technology in the optimal manner is far greater than its misuse.
What is not helpful (at all) is to revert to the baseline and dismiss timely cautionary concerns.
That's a pretty stupid take. A "useful tool" is only useful in a context, and we really only want that tool used to to good. It's not good to have tool used to do bad things. If a school system can be screwed up merely by giving all students free access to guns, maybe is wasn't that great in the first place, amirite?
Essays (for instance) are a means to an end of developing young minds (which is a good thing) by getting them to engage with something. If your "useful tool" makes bypassing that engagement (cheating) simple and easy, those young minds will float through less developed.
There's a lot of hype about LLM tutors, but I don't see that solving much. They'll end up tutoring LLM students, so the real student can invest more effort in playing some addictive video game.
And then that decayed state will become the new normal, and some idiot will declare it's the best because that's what he knows.
The fact that you can cheat in a game does not mean the game itself is flawed.
If you use tools to effectively ghost write your assignments you are not learning.
1) Parents that walk their kids through the homework have more stress not better performing students
2) individualized learning programs exist since the 70s, they were not a substitute - you have to explain something to somebody to fully grasps something - I do not see how anybody wants to explan something to an ai
3) Nearly everybody has been to a school so everybody is an expert
One of the strong areas of school is the in-person-interaction part of learning that can not be easily replaced by ai.
We've had edu-tech ventures for like 25 years now, and they consistently fail. Do you know why? It's not rocket science and it doesn't require some deep analysis.
Children are humans, and they're going to listen to other humans to shape their behavior. Not a screen. Not a program. Not AI.
Just think about it. Suppose you want a well-mannered child. Should you model your manners and surround them with other children who are also learning manners? Or, should you lock them in a room and give them the perfect course in manners from manners.com? Which one will work?
A lot of the best teachers have been driven out of the field, so uninvolved teachers are over-represented. That's a solvable problem, though - good teachers _want_ to teach, we've just built a system that doesn't compensate them.
For my children, my wife and I have never enrolled them in a government school. We call ourselves "unschoolers", but that community probably wouldn't accept us as we'd be unable to pass their implicit ideological purity tests :)
Basically, our kids live alongside us. When they were younger, I'd build them mechanical and electronic toys. That lead to them wanting to help, which lead to some light electrical engineering. When they were old enough to show interest, they got devices of their own. As they got older, that went from an Amazon Fire tablet in a foam protective case to an iPad Pro in a keyboard folio.
We don't filter anything. I log everything at the router level, and have MitM SSL termination configured there. That used to be monitored by a cron job and a set of grep scripts that look for keywords, but these days it's augmented by an AI that summarizes usage and sends me an email once a week. I very rarely look at it.
We have two daughters, and our oldest is 16. There have been a handful of PornHub visits in the past, as is to be expected. The first time it happened, she was 12. I sat down with her privately and asked if she visited it intentionally. She had, because a friend had mentioned an act that she hadn't heard of - so she looked it up. The visit was ~5m long, and that was consistent, so that was the end of the discussion.
Subsequent visits have been similar things, so I only mentioned it in passing to her to let her know it was visible. A longer visit happened a few months ago, shortly after she turned 16. It was still fairly short, and based on the titles and durations it was clear she was looking for information rather than self-gratification. Still, she's shown no reason for me not to trust her and was getting to the age where privacy should be a growing concern for her.
I pulled her aside again, only this time I asked her to bring me her devices. I told her that she's shown that she can be trusted, and that while she is definitely still a child her parents trust her to make good decisions for herself. I excluded her devices from the router logs, removed the MitM SSL certs, and showed her how it was all set up.
Our oldest taught herself to read at a high school level when she was seven years old, because she needed to be able to read quest text to progress in Guild Wars 2.
My youngest taught herself arithmetic with fractions because she loves to cook, and wanted to be adjust recipes to make batches large enough to sell at the farmer's market.
Both of my children learned algebra, geometry, and some basic trig on road trips. When I'm driving we often just talk through things. Sometimes that's math, sometimes history, sometimes politics. Both of them have gone from basic arithmetic to polynomials over the course of a single road trip (different trips!). For geometry and trig, we used iPads and Apple Pencils. These days I carry a reMarkable Paper Pro that gets used a lot for similar reasons.
---
All of the above is to say: kids learn thing naturally. Teachers are not a requirement for learning, nor is school. If you give a child a supportive environment and access to the resources they need to support them, they will seek out knowledge.
But now we have AI, which means that these people, who lack appreciation for the value of these classes and consequently don't even really understand what it is they lack, are better able to foist their beliefs in the unimportance of a well-rounded education on the rest of us.
I will say, it was CIS classes in college which got me to expand on my writing, not English classes. In high school, even with English class every year, I tried to do the bare minimum. It wasn’t until the papers were required to be 25-30 pages long that I was forced to really flesh things out into a narrative, rather that what amounted to a bulleted list shoved into a paragraph.
This was a doubled edged sword, as I now find it much more difficult to be concise in my writing, so that takes a focused effort.
There is also a lot to be said for finding the intersection between technology and X. Having a well rounded education can provide more opportunities for people to discover where these intersections can foster new growth or innovation. I often wish I went into a non-computer field, so I could have applied technology to the job, rather than technology feeling like the job itself.